lkml.org 
[lkml]   [2006]   [Apr]   [5]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
Subject[patch 2/3] mm: An enhancement of OVERCOMMIT_GUESS
This patch is the main part of this enhancement.
Additionally, the patch includes better error handling in __vm_enough_memory().

---
Hideo Aoki, Hitachi Computer Products (America) Inc.
This patch is an enhancement of OVERCOMMIT_GUESS algorithm in
__vm_enough_memory() in mm/mmap.c.

When the OVERCOMMIT_GUESS algorithm calculates the number of free
pages, the algorithm subtracts the number of reserved pages from
the result nr_free_pages().

Signed-off-by: Hideo Aoki <haoki@redhat.com>
---

mmap.c | 18 +++++++++++++++---
1 files changed, 15 insertions(+), 3 deletions(-)

diff -purN linux-2.6.17-rc1-mm1/mm/mmap.c linux-2.6.17-rc1-mm1-idea6/mm/mmap.c
--- linux-2.6.17-rc1-mm1/mm/mmap.c 2006-04-04 10:43:57.000000000 -0400
+++ linux-2.6.17-rc1-mm1-idea6/mm/mmap.c 2006-04-04 14:56:51.000000000 -0400
@@ -121,14 +121,26 @@ int __vm_enough_memory(long pages, int c
* only call if we're about to fail.
*/
n = nr_free_pages();
+
+ /*
+ * Leave reserved pages. The pages are not for anonymous pages.
+ */
+ if (n <= totalreserve_pages)
+ goto error;
+ else
+ n -= totalreserve_pages;
+
+ /*
+ * Leave the last 3% for root
+ */
if (!cap_sys_admin)
n -= n / 32;
free += n;

if (free > pages)
return 0;
- vm_unacct_memory(pages);
- return -ENOMEM;
+
+ goto error;
}

allowed = (totalram_pages - hugetlb_total_pages())
@@ -150,7 +162,7 @@ int __vm_enough_memory(long pages, int c
*/
if (atomic_read(&vm_committed_space) < (long)allowed)
return 0;
-
+error:
vm_unacct_memory(pages);

return -ENOMEM;
\
 
 \ /
  Last update: 2006-04-06 01:50    [W:0.055 / U:0.048 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site